Blackstone Is Selling Picks and Shovels Into the AI Boom — and Getting a Lower Multiple for It

Earnings power accelerated 36% in Q3 while the shares sit 40% below their 2024 peak; management's answer is to hug the one part of the AI stack with 20-year contracts attached

BX · Earnings Call · 2026-10-22

The disconnect, marked to market

Blackstone's third quarter was, on the numbers, a very good quarter. Total revenue rose 36% year over year to $5.0B and 39% sequentially, and net income climbed 45% to $2.4B. Interest coverage sits at 19.4x, nearly double where it was a year ago. By any operating measure, the firm is accelerating. The stock is not. BX is 40.3% below its November 2024 peak of $199.05, 20.5% below its August 13 high, and down 7.1% over the last ninety sessions — a fresh leg down going into this print, not a stale one. Price to net income has compressed to 13.0x, down 33% year over year. Stephen Schwarzman put the tension plainly: “Over the past 2 years, while our shares are down meaningfully, our earnings have grown nearly 50%. With commensurate growth in our dividend.” — Stephen Allen Schwarzman, Chairman and Chief Executive Officer · 2026-10-22 That is the whole story in one line: a de-rating, not a deceleration. What matters for a reader today is whether the de-rating is about AI concentration risk — the question analysts asked four separate ways on this call — or about the real estate segment that still refuses to inflect.

"Long term contracted" is now the entire thesis

Watch what the language did. For several quarters Blackstone's AI story was told in the vocabulary of opportunity: AI infrastructure, shortage of compute, exposure to data centers. This quarter, the firm's single highest-ranked keyword is long term contracted — a defensive, structural phrase, not an offensive one. When the top of the keyword table shifts from what you own to how you own it, management is answering a bear case rather than pitching a bull case. And the bear case was loud. Asked about AI financing risk, Jonathan Gray leaned entirely on counterparty structure: “the fact that we wait till we have these long term contracts before we put the vast majority of capital on the ground that seems to be a very disciplined approach to deploying capital.” — Jonathan D. Gray CIMA, President and Chief Operating Officer · 2026-10-22 On concentration, he quantified it: roughly 7% of $1.4 trillion of AUM touches AI, and “the vast majority of that 7% is in data centers that have 15 to 20 year lease terms at initiation with 96% investment grade counterparties” — Jonathan D. Gray CIMA, President and Chief Operating Officer · 2026-10-22 averaging a $2.5 trillion market cap. Here is the contrast that makes this interesting. On the market's own tape, AI data center was one of the strongest themes of the trailing year — and is now one of the largest 30- and 90-day decliners, with 48 tickers down against 6 up. The same is true of data center capacity. Blackstone is leaning harder into the theme precisely as the tape rotates out of it. Its defence is that the physical bottleneck has not loosened: compute prices “are actually up 50% year-on-year” — Jonathan D. Gray CIMA, President and Chief Operating Officer · 2026-10-22 and data-center leasing is running at three times last year's level. Schwarzman's framing was blunt: “Seldom my career have I seen such a supply constrained market.” — Stephen Allen Schwarzman, Chairman and Chief Executive Officer · 2026-10-22 The genuinely novel argument on this call was Gray's — that compute is on the same path aircraft took to becoming an investment-grade asset:

When you look at the airline business, you know, lending against aircraft is an investment grade business, not because Spirit Airlines or an emerging market airline has great credit themselves, but it is because of the shortage of commercial aviation in the world. Today, when it comes to compute, be it data centers or chips, people are not putting much value on the underlying assets.
Jonathan D. Gray CIMA, President and Chief Operating Officer · 2026-10-22

That is an argument about future collateral recognition, i.e. a case for spreads to compress, not a case about today's cash flows. Worth flagging as a claim, not a fact.

The one hard, checkable new number

Buried in the Q&A was the most concrete promise of the call. Michael Chae said that “what we have under contract that is expected to close in terms of carry realizations alone in the fourth quarter is expected to surpass the total net realization levels for the firm from the third quarter.” — Michael S. Chae, Vice Chairman and Chief Financial Officer · 2026-10-22 That is a magnitude, not a metaphor — and it is new. In July, Gray would only say the engine was coming: “here, we have now been in a period of basically 4 years with some similarities... but, ultimately, I think we know where this is heading.” — Jonathan D. Gray CIMA, President and Chief Operating Officer · 2026-07-23 Backing the promise is a net accrued performance revenue receivable of $7.3 billion, roughly $6 per share, up 12% year over year, with nearly a third of the corporate private equity portion already public and liquid. The quieter re-rating engine is BXMA. The Absolute Return complex grew 22% to $114 billion, posted record organic net inflows, and notched a 26th consecutive positive quarter — beating a 60/40 blend by roughly 170 basis points a year since 2020. It is now being productised for individuals, with a new hedge fund vehicle that raised about $400 million out of the gate. Notably, fee rate ranked second among the firm's keywords this quarter, and Chae pointed to a stabilization in average fee rate there — the commingled mix coming back at higher fees. For a business that has spent a decade defending fee compression, that is a small but real change of direction.

What still has not turned

Real estate remains the hole in the story. Opportunistic BREP funds declined 3.2% in the quarter as the firm raised valuation cap rates in response to higher base rates; core-plus was up 1.6%. “We would expect real estate base management fees to remain broadly stable through year end and into 2027.” — Michael S. Chae, Vice Chairman and Chief Financial Officer · 2026-10-22 Stable, not growing — after the July call where Gray called the recovery a question of "when, not if." He did concede a better set of tumblers this time: new supply at a decade-plus low, debt markets open, and data-center leasing at 6 gigawatts this year, triple last year's pace, with data centers and logistics now 54% of the global real estate book. And the 2027 management-fee guide — fee growth in the low double digits — has now been repeated for a third consecutive quarter, and Chae confirmed it is inclusive of catch-up fees rather than a clean core number. Recurring guidance is not a problem; recurring guidance that keeps slipping a year out is. The fourth quarter is where the claim finally gets tested in cash. The bet you are being asked to underwrite is simple: 13 times earnings for a business compounding revenue 36%, where nearly half of performance-revenue-eligible AUM needs no asset sale to pay out. The risk is equally simple — the AI trade has already begun marking down in the tape, concentration gets asked every quarter now, and the segment that would re-rate the multiple is still marking time.